Welcome to NYU 2015 Conference on Digital Big Data, Smart Life, Mobile Marketing Analytics Conference Co-chairs: Xueming Luo GBM Center Director/Founder Professor of Marketing, Strategy and MIS Fox School of Business, Temple University Russ Winer William H. Joyce Professor of Marketing Stern School of Business New York University 10/23/2015 1
2015 Big Data Conferences Munich, Shanghai, Chicago Oct 23 at NYU Stern Conf themes Industry Academic 2
The World Has Changed so is Customer Journey 50s 60s 70s 80s 90s 00s 10s Consumer Buying Behavior: Process Satisfaction and Loyalty Services CRM Market Ori Scanner data CLV 1-to-1 UGC data Mobile/ IoT Social media Copyright by GBM Global Center for Big Data in Mobile Analytics, All rights reserved
Mobile = On-the-Go, smart IoT Smart Home Office Smart Wearables Copyright by GBM Global Center for Big Data in Mobile Analytics, All rights reserved Smartphones 4
Omnichannel (offline - Mobile - online) Customer Journey 0-moment Mobile SEO In-store SMS Coupon App Retargeting Mobile Share Hear from a friend at work, pull out iphone and google Store locator, call, QR, m-mapping Try in store Compare prices at rival stores SMS opt-in $25 off App alerts Mobile Wallet Facebook likes Copyright by GBM Global Center for Big Data in Mobile Analytics, All rights reserved
What is unique about mobile big data? Personal-always me: precise consumer profiling Who Where Customer Journal Contexts When How Real Time always on: Consumer behavior temporal Location contexts: Physical + digital word Rich interaction: touch, shake, scan, movement, sensors provides marketers an enriched consumer behavior context Copyright by GBM Global Center for Big Data in Mobile Analytics, All rights reserved 6
Mobile Marketing Opportunity, despite Privacy Concerns High opt-in Rate of Mobile Notification by Industry Verticals 7
Mobile Meets MSI Tier 1 Priorities Customer journey, engagement in a multi-media, multiscreen, and multi-channel era? Mobile = physical + digital worlds Hyper-local 24/7 ubiquitous engagement Marketing analytics with big data, real-time decisions, causality of marketing? 2B smartphone users worldwide Real-time pricing (Uber) and social m-commerce (WeChat) 8
Mobile Marketing Opportunity 2 billon people with smartphones worldwide Mobile ad spending $120B by 2018, > TV ad Influence 80% of $2.2T brick-and-mortar sales A hub for physical and digital worlds (OMO) In-store mapping Location services Apps Uber Macy s M-Coupon M- Search M-Social Copyright by GBM Global Center for Big Data in Mobile Analytics, All rights reserved
Mobile Marketing Opportunity
Customer Journey Mobile Context Analytics Research Framework Mobile Marketing Mix Performance Product App Uber Macy s Ad/Coupon Geo-fencing Ibeacon in-store User Journey Smart prediction Disco/Engage/Buy Location Hyper-local Omni-channel Price Freemium App surge price Customer Journey Contexts Targeting any time locat Weather, crowds Competition Geo-conquesting Privacy Personalization Global/Emerg markets Donation Call to action Branding Conversion ROI Loyalty User WOM Network value TV PC Mobile/IoT Copyright by GBM Global Center for Big Data in Mobile Analytics, All rights reserved
Big Data Mobile Analytics Center Research Examples Geo-fencing and lead time: (2014 MgSc) Geo-conquesting: competitor (2015 JMR) Subway crowdedness: mobile immersion (2015 MkSc) Enduring effects of flash deals (2015 ISR) Hyper-competitive pricing (working) Donation self-signaling (working) Omnichannel couponing (working) App pricing and habit (working) Weather and Mobile Promotions: 10-million-users (working) 12
Motivation -- Weather Ads Examples 13
App, Weather, and Demand Mobile App 150 mm Location downloads Effect on in-store Purchasing (ex. Walmart, P&G) Weather Condition 14
Literature on Weather Weather affects 1/3 of GDP, across B2B B2C sectors Restaurants, apparel, automotive, insurance, agriculture (Shah 2013) Shapes everyone s activities and psychology Sunlight produces serotonin, good mood (Lambert et al. 2002) Rain leads to depression and crime (Hsiang, Burke and Miguel 2013) 15
Challenges and Spurious Correlations Cannot manipulate weather Weather confounded with geographic locations New York vs. Florida Deviations and changes in weather No-promotion baseline for incremental ad response Mood explanations Big/small data generalizability 16
Field Data: 10 Million Mobile Users Cannot manipulate weather Quasi-experiment to randomize SMS ad exposures SMS ads promoting a video-streaming service deal Randomized treatment ads (prevention framing Don t miss the opportunity ) vs. control ads ( Dear respected customer ) Holdout group Dependent Variable: Purchase Ad 17
Field Data: 10 Million Mobile Users Cannot manipulate weather Quasi-experiment to randomize SMS ad exposures SMS ads promoting a video-streaming service deal Randomized treatment ads (prevention framing Don t miss the opportunity ) vs. control ads ( Dear respected customer ) Holdout group Dependent Variable: Purchase Ad 18
Ad Copy Regulates the Sunny/Rainy Effects Sunny Rainy + + 0 0 - - Non-Prevention Ads Prevention Ads Non-Prevention Ads Prevention Ads Prevention ads: Don t miss the opportunity Non-Prevention ads: Dear respected customer 19
Table 3 Purchase Rate of Hourly Weather Conditions Purchase Rate Number of Observation Hourly Sunny 0.00022 2643566 Hourly Cloudy 0.00016 2847163 Hourly Rainy 0.00013 2834813 Total 0.00017 8325542 0.99 1.00 Kaplan-Meier survival estimates 0 10 20 30 40 50 analysis time Rainy Sunny Cloudy 20
Passed More Robustness Checks Alternative Measure of Weather Backward and forward looking weather Amount of sunlight Interaction with temperature Accounting for Behavior Weekdays 9 am to 5 pm, hazard model Mobile usage behavior 21
Weather Conclusion and Future Directions Big and small data evidence for sunny/rainy days effects Weather-neutral product, mood explanation Mobile Ads Higher ROI Designing ad copy Future research Forecasts-based ads for online and offline omnichannel sales 22